Self-learning adaptive production systems
Efficient manufacturing requires effective control over production, together with monitoring and control of secondary processes, including energy efficiency and maintenance. Currently, primary control activities and secondary processes are generally separated despite the clear benefits of integrating activities in terms of flexibility, equipment availability and costs. The EU-funded project 'Reliable self-learning production systems based on context aware services' (SELF-LEARNING) developed self-learning solutions for tight and real-time integration of control and maintenance in production systems. The approach was based on principles of service-oriented architecture (SOA) in which individual software units perform specific functions as services to other applications. Different units could then handle control, energy and maintenance, but when interconnected, share information over a network. A major advantage is that human intervention is not needed and this approach has more flexibility, reliability and adaptability. SOA approaches have historically been plagued by problems related to reliability, availability, interoperability, security and trust. Context awareness was exploited in this project to ensure smooth integration of units and appropriate adaptation. Scientists developed context models and an associated context extractor to access parameters related to processes, machinery and conditions for both self-learning and feeding into the adapter module. The latter then provided holistic real-time reaction to conditions. The SOA-based infrastructure included a security and trust framework for safe information flow, data isolation and damage limitation. SELF-LEARNING employed SOA principles together with self-learning and context awareness to facilitate context-specific integrated and harmonised adaptation of production systems. The core technologies are available in open-source format to facilitate broad access and uptake by industry. This software is expected to help resolve critical problems in manufacturing environments by reducing scrap and energy consumption with enhanced efficiency at lower costs.